1 |
On Fri, 2 Jun 2017 02:23:44 -0400 |
2 |
Alan Grimes <ALONZOTG@×××××××.net> wrote: |
3 |
|
4 |
> without spending all day and all night cut-pasting filenames into |
5 |
> another terminal and running rm on them... |
6 |
|
7 |
Looking at the candidate you showed: |
8 |
|
9 |
k3b: |
10 |
|
11 |
version=2.0.3-r5 slot=4 stable |
12 |
version=17.04.1 slot=5 testing |
13 |
|
14 |
It seems like its plausible you recently changed your keywording choices |
15 |
somewhere from "arch" to "~arch", bringing a lot of untested upgrades |
16 |
with it. Or, you allowed portage to perform far more auto-unmasks than |
17 |
really made sense for what you're doing. |
18 |
|
19 |
I could be wrong though. |
20 |
|
21 |
But what appears to be the problem is you have 2 k3b versions, which, |
22 |
according to slots, should be permitted to be co-installed, but |
23 |
according to the files they install, can't be co-installed. |
24 |
|
25 |
This seems like possible cause for opening a bug on k3b, so that |
26 |
k3b:5 blocks against k3b:4 and vice versa. |
27 |
|
28 |
But as for the other cases you saw, I can't really comment, because its |
29 |
seldom the case that multiple packages are having file collisions for |
30 |
the same reason. |
31 |
|
32 |
The only usual reason for that sort of thing happening on a broad scale |
33 |
is /usr/ getting provisioned during install, and staying, but the |
34 |
contents index in /var/db/pkg getting lost somehow due to the segv + |
35 |
reboot, leaving the files there, but leaving portage with no memory of |
36 |
where they came from. |
37 |
|
38 |
I've had that sort of thing happen before, but its very rare. |